Cancer immunotherapy trial design with long-term survivors.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Cancer Research

Background:

• Cancer immunotherapy trials often show improvements in both short-term risk reduction and long-term survival.
• Mixture cure models are suitable for designing such trials, but hazard functions may crossover, violating proportional hazards assumptions.
• Standard log-rank tests (LRT) may lose statistical power for detecting long-term survival benefits when proportional hazards are violated.

Purpose of the Study:

• To propose a change-sign weighted LRT for cancer immunotherapy trial design.
• To develop a sample size formula for the weighted LRT.
• To evaluate the efficiency of the proposed weighted LRT compared to the standard LRT.

Main Methods:

• The study proposes a change-sign weighted log-rank test (LRT).
• A sample size formula for the weighted LRT was derived for trial design.
• Simulation studies were conducted to compare the standard LRT with the change-sign weighted LRT.

Main Results:

• The proposed weighted LRT is designed to detect both short-term risk reduction and long-term survival improvements in cancer immunotherapy.
• Simulation studies assessed the comparative efficiency of the standard and weighted LRTs.
• The weighted LRT is expected to maintain power in detecting long-term survival benefits when hazard functions crossover.

Conclusions:

• The change-sign weighted LRT offers a valuable tool for designing cancer immunotherapy trials.
• This method can improve the detection of long-term survival benefits, especially when proportional hazards assumptions are not met.
• The derived sample size formula aids in robust trial design for immunotherapy studies.